A screen reading system for providing accessibility to mul­timodal document content includes a memory storing: a document file defining a displayable document page con­taining text and images presented in panels to define a graphic narrative; an accessible content database stores audio files for the text and images, defined page locations for the audio files, panel locations for panels, and narrative sequencing for at least some of the audio files and the panels; and a processor-executable screen reading engine that when executed by a processor displays the document page on a display, receives user input to the display, and plays the audio files in response to the user input. The screen reading engine facilitates multiple modes of screen reading and can provide haptic feedback.